Requires version 1.1.12 or later of the Archetype platform.
Overview
This endpoint cancels a pending or running eval: it moves the eval to cancelled, stops its
execution, and returns the updated eval. A predictions artifact written so far is kept, marked
partial.
An eval that has already terminated (with status whose value is completed, failed, or
cancelled) has nothing to cancel and returns HTTP status code 409.
For an eval in a project, you get 403 when you can see the
project but not modify it, and 404 when you cannot see the project.Canceling is the prerequisite for deleting a non-terminal eval. Attempting to delete a
non-terminal eval fails, returning HTTP status 409. Cancel the eval first to avoid this error.
Request
string
required
Eval evl_ id.
Response
Returns the cancelled eval, in the same format that Get
Eval returns.
string
required
TypeID-encoded eval identifier (evl_ prefix).
string
required
Human label for the eval.
string
required
Organization identifier the eval belongs to.
string
required
The blueprint being evaluated.
string
required
The run’s headline, as <target>.<objective>.
object
required
The objectives this run computes, keyed by target name.
boolean
required
Whether the run emits the predictions artifacts alongside the metrics report, one per target.
array
required
The examples the eval was created against, as resolved.
string
required
The eval’s lifecycle status; cancelled after a successful cancel.
string
The project the eval is in: its blueprint’s. null outside any
project. v2.0.0+
string
required
Subject id (usr_... or key_...) that created this eval.
string
required
Creation timestamp (date-time).
string
When scoring began. Absent if the eval was cancelled while still pending.
string
When the eval reached a terminal status.
object
Always absent on a cancelled eval, since the report is populated only when an eval reaches
completed.
array
References to the files the run produced. A cancelled eval keeps its partial predictions entry — the
rows scored so far are still readable.
A cancelled eval keeps the rows it already scored. Its predictions artifact stays in
output_artifacts with metadata.status set to partial.
string
Why the eval failed. Present only when status is failed, so absent on a cancelled eval.
curl -X POST "$ATAI_API_ENDPOINT/agents/evals/evl_01jcb0h2m6t4xr9nv3k7pdzs5y/cancel" \ -H "Authorization: Bearer $ATAI_API_KEY"